An MCP server that bridges any AI assistant (Claude Desktop, Cursor, etc.) to the StellarPay402 marketplace. The AI discovers paid APIs as MCP tools and autonomously pays for them in USDC on Stellar testnet — no human in the loop.
Single tool 'dynamic_marketplace_tools' has critical definition gaps. Tool name lacks action verb (should be 'call_marketplace_api' or 'invoke_endpoint'). Description is generic and does not explain WHAT the tool does operationally, it describes the marketplace concept rather than the tool's behavior. Input schema present but minimal: only 'path' and 'query' strings with basic descriptions. No output schema documented. No error handling guidance. No parameter constraints (enums, ranges, formats). Tool appears to be a catch-all proxy rather than a well-defined, composable tool. Per Arcade pattern:tool, tool names must start with action verbs and descriptions must state WHAT, WHEN, and prerequisites. This tool fails both.
Dynamically generated tools from the StellarPay402 marketplace. Each endpoint in the marketplace becomes an MCP tool with the pattern {userSlug}_{slug}. Tools are named alphanumerically and include pricing, provider info, and auto-payment via x402 on Stellar testnet.
Tool name 'dynamic_marketplace_tools' lacks action verb. Should be 'call_marketplace_endpoint', 'invoke_api', or 'query_marketplace'. LLMs infer intent from verb-noun naming; this name is a noun phrase that does not signal an action.
Tool description is conceptual, not operational. It explains the marketplace and x402 payments but does not state: What does THIS tool do when called? What parameters does it accept? What does it return? LLMs cannot determine when to select this tool.
No output schema documented. LLMs cannot plan downstream calls or extract results. What fields does the response contain? Is it JSON, plain text, or raw API response? Unknown.
Input parameters 'path' and 'query' are generic strings with no constraints, enums, or format guidance. LLMs will hallucinate arbitrary values. No validation rules, length limits, or examples of valid inputs.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 23 | 2026-07-28+ | v2 |
No error handling guidance. If the tool fails (invalid endpoint, payment rejected, timeout), what should the LLM do? Retry? Ask the user? Call a different tool? No recovery path documented.